Species interactions are key factors determining the distribution and structure of species assemblages. Owing to their central positions, mid-ranking mammalian carnivores are involved in interactions with numerous species, including competition for resources and instances of killing by higher ranking predators. Lethal interactions can directly influence species’ demography. However, the fear of lethal interactions, competition or both may also affect when and where individuals are active (i.e. non-lethal interactions). Although differences in body–size and trophic overlap are known predictors of the frequency of lethal interactions, their influence on non-lethal interactions is uncertain. Through camera trapping, we studied non-lethal interactions between a small mesocarnivore (pine marten), a potential killer and intense competitor (red fox) and a moderate competitor and unlikely killer (Eurasian badger). We determined overlap and differences in their diel activity patterns and the degree of spatial overlap in two seasons with contrasting resource availability. Additionally, we estimated the effect of larger carnivore detection rates on pine marten detection rates. We also compared time intervals between pine marten visits to baited stations in the absence and presence of either or both larger carnivores. Our results are consistent with pine martens distributing their daily activity to maximize overlap with prey and to minimize competition and risk of aggression over the spatial scale. Pine martens also responded to the immediate threat of larger carnivores irrespective of the threat they pose by taking 4–7 days longer to re-visit a station. Small-scale non-lethal interactions such as these may enable pine martens to coexist closely with two larger carnivores, yet it remains uncertain whether their population incurs a demographic cost through restricted access to resources. Carnivore's risk-avoidance strategies could be harnessed to protect prey species of interest. However, our results suggest avoidance is short-lived and recurrent stimuli would be necessary.